dream-memory
Dream Memory
Use this skill when you want a deliberate memory-consolidation pass instead of storing more raw notes.
This bundle is intentionally portable. It borrows the workflow shape from Claude Code's dream system, but it does not depend on Claude Code internals.
Use It For
- nightly or manual memory cleanup
- merging overlapping memory notes
- normalizing relative dates into absolute dates
- pruning stale or contradicted memories
- keeping
MEMORY.mdsmall and index-like
Avoid It For
- saving a single new memory from the current turn
- persisting code facts that should be re-read from source
- replacing plans, task boards, or issue trackers
Quick Start
Inspect a memory directory before running the consolidation prompt:
python3 {baseDir}/scripts/dream_memory.py --memory-root /path/to/memory --transcripts-dir /path/to/transcripts
Then use the portable prompt in references/prompt-template.md.
Workflow
- Inspect
MEMORY.md, topic files, and recent logs or transcripts. - Identify durable facts worth keeping.
- Merge into topic files instead of creating near-duplicates.
- Remove stale or contradicted memory.
- Rewrite
MEMORY.mdas a concise index.
Rules
MEMORY.mdis an index, not a content dump.- Prefer topic merges over new file creation.
- Convert relative dates to absolute dates.
- Do not store code-state facts that can drift.
Supporting Files
- Prompt template: references/prompt-template.md
- Source notes: references/source-notes.md
- Helper script:
python3 {baseDir}/scripts/dream_memory.py ...
More from learnprompt/cc-harness-skills
memory-extractor
Extract durable memories from recent conversation turns into user, feedback, project, and reference categories while avoiding stale code-state facts.
23verification-gate
Run a read-only verification pass after implementation to check whether completion claims are real, validation actually ran, and obvious edge cases or regressions were missed.
23structured-context-compressor
Compress a long agent conversation into a nine-part continuation summary that preserves request, files, errors, user messages, current work, and the next aligned step.
19swarm-coordinator
Coordinate multiple agents by splitting work into research, synthesis, implementation, and verification, assigning ownership, and keeping the coordinator focused on integration rather than raw exploration.
14kairos-lite
Build a lightweight proactive mode with scheduled checks, sleep intervals, concise user briefs, and expiry safeguards so an agent can work in the background without becoming an uncontrolled daemon.
13